Distribution and molecular phylogeny of biliary trematodes (Opisthorchiidae) infecting native Lutra lutra and alien Neovison vison across Europe

a School of Biosciences, Cardiff University, Museum Avenue, Cardiff CF10 3AX, UK b Department of Infectious Disease Epidemiology, Imperial College London, Norfolk Place, London W2 1PG, UK c Department of Bioscience, Aarhus University, Grenåvej 14, Kalø, DK-8410, Rønde, Denmark d Wildlife Veterinary Investigation Centre, Chacewater, Cornwall TR4 8PB, UK e Norwegian Institute for Nature Research, PO Box 5685, Sluppen, NO-7485 Trondheim, Norway f Groupe Mammalogique Breton, Maison de la Rivière, 29450 Sizun, France g Department of Contaminant Research and Monitoring, Swedish Museum of Natural History, PO Box 50007, SE-104 05, Stockholm, Sweden h VetAgro Sup, Campus Vétérinaire de Lyon, 1 Avenue Bourgelat, 69280 Marcy-l'etoile, France i ALKA Wildlife, o.p.s., Lidéřovice 62, CZ-380 01 Peč, Czech Republic j Charles University in Prague, Third Faculty of Medicine, Ruská 87, CZ-100 00 Prague 10, Czech Republic


Introduction
Although parasites play an integral role in ecosystem functioning [1][2], there is incomplete knowledge of their geographic and host ranges. This arises in part because of the morphologically cryptic nature of many parasitic taxa and, specifically, the challenge of detection and species determination [2][3]. This is particularly problematic for invading parasites, which often present serious risks to novel host populations, largely because of naïve host immune responses coupled with disruption of ecosystem equilibrium [4]. The recent identification of Pseudamphistomum truncatum (Rudolphi, 1819) and Metorchis albidus (Braun, 1893) Loos, 1899 (Trematoda: Opisthorchiidae) in Britain [5][6] caused initial concern because of the biliary damage to otters that was associated with both digeneans, and speculation over their alien status in Britain [5][6][7].
P. truncatum occurs across Europe and was reported in the early 1900s [8] from mammals that are native to Britain. However, without details of host origin, the recorded hosts (red fox Vulpes vulpes [also see 7], grey seal Halichoerus gryphus, domestic cats Felis domestica and dogs Canis familiaris, common seal Phoca vitulina and the harp seal Phoca groenlandica [8]) may have been sampled in continental Europe. Therefore, it remains unclear whether P. truncatum is a recent invader in Britain.
Historically, the taxonomy of the Metorchis genera has been complicated because of variable parasite morphology in multiple vertebrate

Contents lists available at ScienceDirect
Parasitology International j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / p a r i n t hosts (including man) across Eurasia [9][10][11][12].  [12] and references therein]. There are only ambiguous historic records of M. bilis/M. albidus in Britain [8], whereas M. crassiusculus has been noted with more conviction from British birds but only using morphological identification methods [13]. Species determination of a morphologically challenging genus like Metorchis can now be addressed molecularly [e.g. [14], Heneberg, pers. Comm.]. The damage associated with host-parasite interactions makes the identification of parasitic species and their geographic ranges particularly crucial for successful conservation efforts [15][16].
The life-cycle of the Opisthorchiidae family has been elucidated only rarely [see [12]] but involves two intermediate and a definitive host. For both P. truncatum and M. bilis, the first intermediate hosts are freshwater snails. A free-living cercarial stage is then released and encysts in a freshwater fish intermediate host. The development of P. truncatum can be completed once consumed by a mammalian definitive host, whereas M. bilis seems more generalist in terms of suitable definitive hosts (piscivorous mammals and birds) [8][9][10][11][12][13].
Here, we aimed to explore the distribution, intensity and molecular phylogenies of P. truncatum and M. bilis in otters and mink from across Europe. To address this, we collected samples across Europe, confirmed species identification using internal transcribed spacer region II (ITS2) ribosomal DNA sequences, and compared genetic diversity between the parasite species and between populations using two mitochondrial DNA markers (COX1 and COX3).

Sample collection
Gall bladders of 723 Eurasian otters (L. lutra) and 144 American mink (N. vison) were sourced from across Europe. Samples were included from Britain, Czech Republic, Denmark, France, Germany, Norway, Poland, Scotland, and Sweden, and preserved in 95% ethanol (a list of samples and host locations is provided in Supplementary Information 1). Each gall bladder was opened along its length in a petri dish containing fresh ethanol. The bladders were examined thoroughly and then rinsed, everted and the mucosa was finely combed to ensure all parasites were found, morphologically identified using a dissecting microscope according to [17] and counted. A sub-sample of parasites (N = 65, see Table 1) was selected for molecular analysis. For British samples (positive cases were found within England and Wales), stratified random sampling was applied, to select trematodes broadly representative of the geographic distribution previously identified [6] (P. truncatum from the Counties of Somerset, Dorset, Gwent and Powys; M. bilis from Bedfordshire, Cambridgeshire, Essex, Hertfordshire and Suffolk). For continental samples DNA sequencing was performed on all samples. A general linear model (GLM), with an associated binary error distribution, was used to compare the parasite prevalence (the infection status of an individual regardless of the number of parasites present; infected = 1, uninfected = 0) between European regions where the sample size was equal to or larger than 10. To investigate intensity (the number of parasites infecting each individual, excluding those without infection) differences across host populations in Europe, a GLM (with a negative binomial error distribution) was fitted to the intensity data for regions where sample size of infected hosts was greater than or equal to 4. This analysis of intensity was limited however because of the small sample size available ( Table 1). All analyses were conducted in R, version 3.2.0 [18].

DNA analysis
The internal transcribed spacer II (ITS2 P. truncatum: GenBank Accession number: JF710315; M. albidus (synonymous with M. bilis): GenBank Accession number: JF710316) of the ribosomal DNA region was used for species discrimination because it has been shown to be relatively conserved in the Digenea and has been used previously for interspecific analyses [12,14,[19][20] while fragments of the mtDNA COX1 and COX3 genes were used to examine genetic variation of P. truncatum and M. bilis populations across Europe.
DNA was extracted from whole individuals. The tissues were stored in 90% ethanol, which was evaporated by gentle heating (55C) prior to digestion of the sample. Whole individual trematodes were digested for 3 h at 55°C in 15 μl TE buffer containing 0.45% Tween 20 and 2 μg Proteinase K, followed by 10 min at 95°C to denature the proteinase K [adapted from 21]. This treatment was sufficient to extract the DNA in preparation for the polymerase chain reaction (PCR), the mixture was centrifuged prior to use in the PCR. The PCR was conducted using 2 μl of the DNA extract in a final volume of 10 μl, containing: 1× PCR buffer II (Applied Biosystems), 2 mM MgCl 2 (Applied Biosystems), dNTP (0.25 mM each), primers (1 μM each) (ITS2 rDNA: Ophet F1 5′-CTCG GCTCGTGTGTCGATGA-3′ and Ophet R1 5′-GCATGCARTTCAGCGGGTA-3′ see [22]; or COX1 mtDNA: ThaenCO1F 5′-CGGGTTTTGGAGCGTCATTC-3′ and ThaenCO1R 5′-ACAGGCCACCACCAAATCAT-3′; or COX3 mtDNA: CO3FTremat 5′-ATGAGWTGATTACCKTT-3′ and CO3RTremat 5′-ACAACC Table 1 Summary of otter (Lutra lutra) and mink (Neovison vison) biliary parasites. The country of origin (location), total number of hosts examined (N), number of infected hosts and the corresponding percentage and intensities of the two species (Pseudamphistomum truncatum and Metorchis bilis) isolated from European piscivorous mammals.  The species present were identified using the ITS2 sequences. Unique COX1 and COX3 haplotypes were identified as follows, and have been assigned GenBank Accession numbers KP869069-KP869078, KP869080-KP869096 (Supplementary information 2): seven P. truncatum COX1 (unique haplotypes from the Czech Republic, France, Germany, Poland, Sweden, 1 unique haplotype from England and Wales, and 1 haplotype common to England and Wales, Denmark and Sweden); eight P. truncatum COX3 (unique haplotypes from the Czech Republic, Denmark, England and Wales, Poland, Sweden, two unique sequences from Germany, and one haplotype common to Denmark, England and Wales, Germany and Sweden); nine M. bilis COX1 (unique haplotypes from the Czech Republic, Denmark, England and Wales, and Germany, three unique haplotypes from France, one haplotype common to Denmark and England and Wales, and one haplotype common to Denmark and Sweden); three M. bilis COX3 (a unique haplotype to England and Wales, one common to Denmark and France, and one common to Denmark and Germany). Bayesian inference (BI) methods were used to reconstruct the phylogenetic relationships among the mtDNA haplotypes for COX1 and COX3 separately for each species using MrBayes version 3.2 [24]. One million Markov Chain Monte Carlo (mcmc) generations were carried out with the initial 25% discarded as burn in. Convergence was assessed through effective sample size values and correlation plots. MrModeltest version 2 [25] was used to estimate the adequate model of sequence evolution of these datasets. For both loci in each species the inferred model of evolution was Hasegawa, Kishino and Yano. We used a gamma-shaped rate variation with a proportion of invariable sites (Invgamma). The human liver fluke Clonorchis sinensis (Trematoda: Opisthorchiidae) was used as an outgroup for both datasets. To complement the phylogenies, and to further explore mtDNA structure, we also constructed haplotype

Table 2
Summary of the data on the genetic diversity of Pseudamphistomum truncatum and Metorchis bilis across Europe: Samples size (number of parasite sequences), the gene sequenced, the population (region of origin), the haplotype diversity, the nucleotide diversity (π), and the number of haplotypes within the population.

Results
In total, 723 otters and 144 mink gall bladders were dissected from samples taken across 8 European countries (see Table 1, Fig. 1). The highest prevalence of P. truncatum was detected in otters from Germany (73%, although the sample size was low; 8 out of 11 otters were infected) (GLM binomial error distribution:  (Table 1). Only mink samples were examined from Scotland and none were infected whilst mink in England had biliary trematodes.
There was no significant difference in the intensity of P. truncatum among infected otters from England and Wales or Germany (the only countries with large enough sample sizes to compare statistically, GLM negative binomial error distribution: F = 0.3167, df = 86, p = 0.57). Equally, the intensity of M. bilis infection did not differ between comparable data sets from France, Sweden, and England and Wales (GLM negative binomial error distribution: F = 2.42, df = 62, p = 0.097).
Only a single ITS2 haplotype was identified across Europe for each parasite species. Nucleotide diversity between the two species was 0.01347 across 300 bp. Analysis of mitochondrial DNA markers COX1 and COX3 showed greater diversity within M. bilis than P. truncatum ( Table 2).

Discussion
The discoveries of both P. truncatum and M. bilis in British mammals occurred as a direct result of systematic screening [5][6]. It is only with systematic and widespread screening (e.g. [28,29]) combined with molecular analysis that taxonomic confusion, such as that surrounding Metorchis [9][10][11][12], can be elucidated, providing a clearer understanding of parasite fauna and disease.
The P. truncatum samples showed low mitochondrial genetic diversity relative to M. bilis. This lower genetic diversity is observed in the shorter branch lengths of the P. truncatum phylogeny, lower haplotype diversity and shorter distance between the haplotypes in the genetic network. These differences indicate considerable differences in the demographic and evolutionary history of these species in Europe. Shared haplotypes across multiple European countries suggests genetic mixing of both parasite species, across Europe. There is insufficient evidence to attempt to date an initial introduction to Britain for either species, but the presence of more than one haplotype (for both species) may indicate more than one introduction event or long-term residency. Shared haplotypes between Scandinavian and British samples (P. truncatum: Hap01, COX1; M. bilis, Hap08, COX1) implicates Scandinavia as a potential origin for both species into Britain.
The branch lengths for the M. bilis phylogeny are an order of magnitude longer than for the P. truncatum phylogeny. This, coupled with the Each black dot on the network corresponds to a single mutational change. There were no meaningful branch lengths for the phylogeny due to low genetic variation at this locus for P. truncatum COX3. difference in the distributions of the haplotypes between the two species implies differing demographic histories, which may correspond to differences in their definitive hosts [e.g., 5,6,8,12,13,30]. Pseudamphistomum truncatum is predominantly reported from mammals [5][6]8] whilst M. bilis (and more specifically synonym M. crassiusculus) are also noted from migratory birds [ [13], Heneberg, pers. Comm.], perhaps increasing the potential for genetic mixing for M. bilis across Europe. Legislation is operative to protect fish from disease (e.g. EU Council Directive 2006/88/EC) but does not apply to most digeneans. In part, this relaxed approach to screening fish for digeneans stems from their reported low level impact on fish [but see [31][32] and it is therefore deemed unnecessary to restrict fish movements on this basis. Consequently, the widespread translocation of fish stocks, natural migration of definitive hosts and fish, alongside movement of snails and parasite eggs with plants, gravel or water, across Europe almost certainly contributes to a widespread distribution of digenean species. For example, P. truncatum is found in Ireland [33] where cyprinid fish (the second intermediate host for both P. truncatum and M. bilis) are not native but were introduced in the 17th Century and have been continually relocated to new habitats across Ireland ever since [34].
The distribution of P. truncatum across Europe appears to be focal, although more extensive sampling would be required to examine this further. The German state of Saxony has a relatively high proportion of P. truncatum infections (8/11) but only 11 otters were examined, whilst the parasite was not found in mink from Scotland (n = 40) or otters from Norway (n = 21). Previously, no biliary parasites were found in another eleven otters screened in Scotland [7]. There are no apparent barriers to spread according to the underlying distribution of some of the potential host species: both the first and second intermediate hosts (gastropod families Lymnaeidae and Bithyniidae, and freshwater fish family Cyprinidae, respectively [ [30], Sherrard-Smith et al., unpublished data]), occur throughout Europe. The predicted warmer conditions across Europe with climate change may suit both species and encourage a Northerly movement of their current distributions [22]. The density-dependent processes that determine the distribution of host populations can result in spatial aggregation of parasite populations [35] and have been used to explain the co-existence of species [e.g. [36][37]. The presence of two biliary parasites in the otter population across Europe may contribute to the observed patchiness in the distribution of each digenean. Speculatively, this may indicate that interspecific competition or some level of acquired host immunity is acting to separate P. truncatum and M. bilis and it is noteworthy that across the entire study, only 5 co-infections (where 8 would be expected by chance; 70/723 P. truncatum infections among 79 M. bilis positive hosts) were observed; 2 from Britain (out of 586 otters) and 3 in Denmark (out of 52 otters), but none elsewhere, of 229 otters). Specifically, a distinction between the geographic distributions of M. bilis and P. truncatum was observed in England and Wales [6], and France, with P. truncatum only found in the Poitou-Charentes Region (a single specimen), and M. bilis only in Brittany (although our sample size in France is relatively small: N = 22 Brittany, N = 19 Poitou-Charentes Region). Regional differences in host diet [e.g. 38] may contribute to geographic variation in parasite exposure. Co-existence of P. truncatum and M. bilis in the same host is rare despite geographic overlap across continental Europe but also in England and Wales where sample size is large enough (n = 586) to make stronger conclusions [6].
The presence of widespread COX1 and COX3 haplotypes, particularly for M. bilis, indicates population mixing throughout Europe. The current study provides an insight into the genetic structure, but also geographic heterogeneity, of two widespread digeneans of threatened wild mammals.